Releases: ibcn-cloudlet/dianne
Releases · ibcn-cloudlet/dianne
DIANNE v0.4.0
Features:
- Major revision of Tensor library resulting in serious performance gains
- No more Java Tensor implementation in favour of simpler code
- Change build system from ant to gradle
- Fixed issues with native compilation not being performed correctly, especially with cutorch
- Improved performance for minibatches
- Improved memory footprint by reusing tensors where possible & triggering garbage collection
- Added various new modules, a.o. FullConvolution, BatchNormalization, ...
- Added various new composite modules, a.o. Inception, Residual, ...
- Extended Convolution to include temporal & volumetric operation
- MaxPooling has indices as optional second output, which can be passed to new MaxUnpooling module
- More efficient loading of datasets
- Added descriptors for datasets, allowing to define custom datasets without coding
- Added dataset adapters for data augmentation, a.o. flip, rotate, crop, ...
- Separated evaluator into generic & classification
- Added more stop criteria for a learn job based on error
- Added options for learning rate decay in SGD
- Perform validation while learning
- Automatically save best weights based on validation error
- Better error handling for jobs (especially on NaN)
- Easier resubmission for jobs, incl. changing parameters
- Fixed issue with URLs redirecting indefinitely
- UI improvements, such as ability to remove modules, job-specific URLs, etc.
- Extended JSON RPC
- Clean up of run configurations
- Various bug fixes
- Updated documentation to reflect new features in more detail
- Added contributor license agreement
DIANNE v0.3.0
Features:
- Major refactoring of learn API
- Added coordinator service for centralized management of tasks (learn, evaluate, act)
- Added dashboard UI for easy submission and follow up of said tasks
- Added JSON RPC for external systems to submit tasks
- Neural networks as composite modules
- Support for recurrent neural nets with a memory module
- Util API for easy programmatic construction of networks
- Improved test setup
- Increased performance for gradient operations on GPU
- Changed build system to gradle
- Separated native libraries into separate bundle fragments
- Improved error handling
- Various small additions and bug fixes
DIANNE v0.2.0
Features:
- DTOs for representing neural networks and their modules
- NeuralNetworkPlatform API as central management API
- NeuralNetwork API facilitating forward/backward calls of neural network
- Processor API for setting up different Trainers
- Distributed training possible synced by Repository
- Reinforcement Learning APIs for modelling Environments, Agents, ...
- Deep Q Learning agent implementation
- Pong RL example
- ALE environment
- DianneCoordinator API for training nn models
- JSON RPC interface to DianneCoordinator
- Add open source license, basic documentation and website
- Various small additions and bug fixes
DIANNE v0.1.0
Initial Dianne Release
Supports
- Modular deployment of neural network
- Supports fully connected and convolutional layers
- Supports basic split/fork/join/concat modules
- Save/load neural network modules
- GUI neural network builder
- Tensor library implementations in plain Java, native based on TH (Torch), cuda lib based on THC
- Basic training procedure for simple neural networks